A [18F]Fluoroethoxybenzovesamicol Positron Emission Tomography Study

Received: 24 May 2018 Revised: 7 September 2018 Accepted: 10 September 2018 DOI: 10.1002/cne.24541

RESEARCH ARTICLE

Regional vesicular acetylcholine transporter distribution in human brain: A [18F]fluoroethoxybenzovesamicol positron emission tomography study

1Neurology Service & GRECC, VAAAHS, Ann Arbor, Michigan Abstract 2Department of Neurology, University of Prior efforts to image cholinergic projections in human brain in vivo had significant technical lim- Michigan, Ann Arbor, Michigan itations. We used the vesicular acetylcholine transporter (VAChT) ligand [18F]fluoroethoxyben- 3University of Michigan Morris K. Udall Center zovesamicol ([18F]FEOBV) and positron emission tomography to determine the regional of Excellence for Research in Parkinson's distribution of VAChT binding sites in normal human brain. We studied 29 subjects (mean age Disease, Ann Arbor, Michigan 47 [range 20–81] years; 18 men; 11 women). [18F]FEOBV binding was highest in striatum, inter- 4Michigan Alzheimer Disease Center, Ann Arbor, Michigan mediate in the amygdala, hippocampal formation, thalamus, rostral brainstem, some cerebellar 18 5Department of Radiology, University of regions, and lower in other regions. Neocortical [ F]FEOBV binding was inhomogeneous with Michigan, Ann Arbor, Michigan relatively high binding in insula, BA24, BA25, BA27, BA28, BA34, BA35, pericentral cortex, and 6Department of Cell and Developmental lowest in BA17–19. Thalamic [18F]FEOBV binding was inhomogeneous with greatest binding in Biology, University of Michigan, Ann Arbor, the lateral geniculate nuclei and relatively high binding in medial and posterior thalamus. Cere- Michigan bellar cortical [18F]FEOBV binding was high in vermis and flocculus, and lower in the lateral cor- 7Department of Psychology, University of tices. Brainstem [18F]FEOBV binding was most prominent at the mesopontine junction, likely Michigan, Ann Arbor, Michigan associated with the pedunculopontine–laterodorsal tegmental complex. Significant [18F]FEOBV Correspondence Roger L. Albin, MD, 5023 BSRB, 109 Zina binding was present throughout the brainstem. Some regions, including the striatum, primary Pitcher Place, Ann Arbor, MI, 48109-2200, sensorimotor cortex, and anterior cingulate cortex exhibited age-related decreases in [18F] USA. FEOBV binding. These results are consistent with prior studies of cholinergic projections in Email: [email protected] other species and prior postmortem human studies. There is a distinctive pattern of human neo- Funding information Michael J. Fox Foundation for Parkinson's cortical VChAT expression. The patterns of thalamic and cerebellar cortical cholinergic terminal Research; NIH-NINDS, Grant/Award Number: distribution are likely unique to humans. Normal aging is associated with regionally specific P01 NS015655P50 NS091856R21 reductions in [18F]FEOBV binding in some cortical regions and the striatum. NS088302; Michael J. Fox Foundation

KEYWORDS acetylcholine, aging, basal forebrain, cerebellum, pedunculopontine nucleus, RRID: SCR_001847, RRID: SCR_007037, striatum, thalamus

1 | INTRODUCTION

Abbreviations: AC, anterior cingulate cortex; ACh, acetylcholine; AChase, acetylcholinesterase; AD, Alzheimer disease; AM, amygdala; BF, basal forebrain; Acetylcholine (ACh) is the primary neurotransmitter of several CNS ChAT, choline acetyltransferase; FEOBV, fluoroethoxybenzovesamicol; FL, floc- neuron populations. These include neurons of the basal forebrain culus; HIP, hippocampal formation; LG, lateral geniculate nucleus; MP, meso- (BF) complex projecting to numerous cortical regions, striatal choliner- pontine; MP4A, N-methyl-4-piperidyl acetate; MVC, medial vestibular complex; PD, Parkinson disease; PET, positron emission tomography; PMP, methyl- gic interneurons, the medial habenula–interpeduncular nucleus pro- 4-piperidinyl propionate; PPN-LDT, pedunculopontine-laterodorsal tegmental jection, projections of the pedunculopontine–laterodorsal tegmental complex; SMC, sensorimotor cortex; SPECT, single photon emission computed (PPN–LDT) complex, the parabigeminal nucleus, some medial vestibu- tomography; SUVR, standardized uptake value ratio; THAL, thalamus; VAChT, vesicular acetylcholine transporter; VER, vermis lar complex (MVC) neurons, and brainstem motor and autonomic

2884 © 2018 Wiley Periodicals, Inc. wileyonlinelibrary.com/journal/cne J Comp Neurol. 2018;526:2884–2897. ALBIN ET AL. 2885 neurons. BF cholinergic pathways participate in the modulation of To advance our understanding of the organization of cholinergic important cognitive functions such as attention and executive func- pathways in humans, we assessed the regional distribution of human tion (Ballinger, Ananth, Talmage, & Role, 2016; Hasselmo & Sarter, brain cholinergic neuron terminals with [18F]fluoroethoxybenzovesmi- 2011; Prado, Janickova, Al-Onaizi, & Prado, 2017). Projections of the col ([18F]FEOBV) positron emission tomography (PET; Aghourian PPN–LDT complex may play important roles in the regulation of sleep et al., 2017; Petrou et al., 2014). FEOBV is a high-affinity ligand for and waking, and some data implicates these neurons in control of bal- the vesicular acetylcholine transporter (VAChT), the protein responsi- ance and gait (Gut & Winn, 2016; Mena-Segovia & Bolam, 2017; Pie- ble for pumping ACh from the cytosol into synaptic vesicles (Prado naar, Vernon, & Winn, 2016). Striatal cholinergic interneurons are the et al., 2017; Prado, Roy, Kolisnyk, Gros, & Prado, 2013). VAChT is subject of intense investigations and play important roles in the con- expressed at uniquely high levels by cholinergic neurons (Arvidsson, trol of a variety of functions mediated by the striatal complex Riedl, Elde, & Meister, 1997). Prior human imaging studies of brain (Deffains & Bergman, 2015; Gonzales & Smith, 2015; Tanimura VAChT expression used less resolute single photon emission com- et al., 2017). puted tomography (SPECT) methods (Kuhl et al., 1994, 1996; Lamare Abnormalities of some of these pathways are implicated in neuro- et al., 2013). Prior PET studies of regional cholinergic terminal distri- logic and psychiatric disease, particularly Alzheimer disease (AD), Par- bution used acetylcholinesterase (AChase) substrates ([11C]methyl- kinson disease (PD), and tobacco abuse. Degeneration of the BF 4-piperidinyl propionate [PMP]; N-[11C]-methyl-4-piperidyl acetate complex is documented well in AD and PD, and degeneration of cho- [MP4A]) to identify cholinergic synapses (Iyo et al., 1997; Kuhl et al., linergic PPN–LDT neurons may drive important aspects of declining 1999). These methods are less specific. AChase, for example, is pro- gait and balance functions in PD and Progressive Supranuclear Palsy duced by midbrain dopaminergic neurons and high regional AChase (Bohnen et al., 2012; Gilman et al., 2010; Hirsch, Graybiel, Duyck- activity was detected in regions, the lateral cerebellar cortices, not aerts, & Javoy-Agid, 1987; Mesulam, 2013; Mufson, Ginsberg, Ikono- suspected to have substantial cholinergic terminal innervation movic, & DeKosky, 2003; Schliebs & Arendt, 2011). Augmenting (Greenfield, 1991; Kuhl et al., 1999). AChase enzyme activity may be cholinergic neurotransmission with acetylcholinesterase inhibitors is regulated by synaptic activity and disease states, a potential obstacle an accepted treatment for dementias, including AD, Lewy body to quantifying regional cholinergic terminal density (DeKosky et al., dementia, and PD-associated dementia. Anti-muscarinic cholinergic 1992). The very high AChase expression characteristic of the striatum agents are used for the treatment of dystonias and for treatment of limits quantification of striatal AChase as PMP and MP4A metabolite PD-associated tremor. Manipulations of cholinergic neurotransmission accumulation may be flow, rather than AChase activity, limited. PMP are being pursued to improve gait and balance functions in PD and MP4A are also, though to a considerably lesser extent, butyryl- (Chung, Lobb, Nutt, & Horak, 2010; Henderson et al., 2016; Kucinski, cholinesterase substrates and butyrylcholinesterase activity changes De Jong, & Sarter, 2017; Li et al., 2015). Nicotinic cholinergic recep- might influence the results of PET studies with these tracers (Snyder tors mediate the addictive effects of nicotine and drugs active at this et al., 2001). receptor family are used to treat nicotine addiction (Anthenelli et al., We describe the regional distribution of [18F]FEOBV binding in 2016). Improved understanding of the organization of cholinergic human volunteers with no evidence of neurologic or psychiatric dis- pathways in the human brain is relevant to efforts to expand useful ease. Our results largely confirm inferences from postmortem human cholinergic pharmacology in the treatment of human neurologic and studies and nonhuman mammal studies but indicate distinctive fea- psychiatric diseases. tures of human cholinergic pathways and indicate effects of aging. Prior investigations in other species, including nonhuman pri- mates, and work with postmortem human tissues provided consider- able information about the organization of human brain cholinergic 2 | METHODS pathways (Albin, Morgan, Higgins, & Frey, 1994; Barmack, Baugh- man, & Eckenstein, 1992; Barmack, Baughman, Eckenstein, & Shojaku, 1992; de Lacalle et al., 1994; De Lacalle, Hersh, & Saper, 1993; Emre, 2.1 | Subjects Heckers, Mash, Geula, & Mesulam, 1993; Fitzpatrick, Diamond, & We studied 29 subjects without histories of the neurologic or psychi- Raczkowski, 1989; Fukushima et al., 2001; Gilmor et al., 1996, 1999; atric disease. We studied 18 men and 11 women. The mean age was Heckers, Geula, & Mesulam, 1992; Ichikawa, Ajiki, Matsuura, & Mis- 47 years (SD = 21 years); age range 20–81 years. No subject was awa, 1997; Jaarsma et al., 1997; Kus et al., 2003; Levey, Hallanger, & using a medication that might affect cholinergic neurotransmission Wainer, 1987; Mahady, Perez, Emerich, Wahlberg, & Mufson, 2017; and none were using tobacco or nicotine products. Informed consent Manaye et al., 1999; Mesulam, 2004; Mesulam & Geula, 1988; Mesu- was obtained from all subjects prior to study entry and this study was lam, Geula, Bothwell, & Hersh, 1989; Mesulam, Hersh, Mash, & Geula, approved by the University of Michigan Medical School IRB. Limited 1992; Mesulam, Mash, Hersh, Bothwell, & Geula, 1992; Roghani, Shir- results were reported previously from subsets of these subjects zadi, Butcher, & Edwards, 1998; Schäfer, Eiden, & Weihe, 1998; Scha- (Albin, Minderovic, & Koeppe, 2017; Petrou et al., 2014). fer, Weihe, Erickson, & Eiden, 1995; Woolf, 1991; Woolf & Butcher, 1985, 1986, 1989; Woolf, Eckenstein, & Butcher, 1984; Zhang, 2.2 | [18F]FEOBV PET Zhou, & Yuan, 2016). Working with postmortem tissues has intrinsic limitations and inferences from studies of other species may be con- [18F]FEOBV was prepared as described previously (Petrou et al., founded by inter-species differences. 2014; Shao et al., 2011). [18F]FEOBV (288–318 MBq) was 2886 ALBIN ET AL. administered via intravenous bolus. Twenty-eight subjects were better visualization of effects of aging. We also used our standard VOI scanned on an ECAT Exact HR+ PET tomograph (Siemens Molecular set developed for Talairach space, which is based entirely on the PET Imaging, Knoxville, TN). One subject was scanned on a Biograph True- scans (no MR), as described previously (Bohnen et al., 2006). Brainstem Point scanner (Model 1094; Siemens Molecular Imaging). Brain imag- parcellation was performed as described previously (Albin et al., 2008). ing was conducted in three imaging periods. The first period began at To complement this regional analysis, cortical and subcortical segmen- injection and continued for 90 min. The subjects were then given a tation was performed with FreeSurfer. The programs were run in the 30-min break, followed by a second imaging period from 120 to fully automated mode. Volumes for all FreeSurfer-defined brain struc- 150 min, an additional break, and then a third imaging period from tures were obtained as part of the standard FreeSurfer output. The 180 to 210 min. Dynamic PET scans were reconstructed using Fourier VOIs were then applied in native MR-space to the co-registered PET rebinning (FORE) and the iterative 2D-OSEM algorithm, with four iter- scans. Results were essentially identical to those of the Neurostat-SPM ations, 16 subsets, and no postreconstruction smoothing. The entire based analysis and are not shown. 18 dynamic sequence was co-registered to the final form of the first To analyze thalamic [ F]FEOBV binding, we used the FreeSurfer- 90-min scanning period. The first seven studies were acquired with a defined thalami, assumed an ellipsoid volume, and estimated the long slightly different protocol, where the first scanning period last for axis of the ellipsoid volume. We then divided each hemisphere of the 120 min, and the second and third periods lasted from 150 to thalamus into octants, based on voxels medial versus lateral, anterior 180 and 210 to 240 min, respectively. versus posterior, and superior versus inferior to the axis.

2.3 | MRI 2.5 | Aging effects

MRI imaging was performed on a 3 Tesla Philips Ingenia System To assess potential aging effects, we took advantage of the fact that (Philips) using a 15-channel head coil. A standard T1 weighted series our study population was divided into two groups; a younger group of a 3D inversion recovery prepared turbo-field echo was performed (N = 13; Mean 25 years; range 20–38) and an older group (N = 16; in the sagittal plane using TR/TE/TI = 9.8/4.6/1041 ms; turbo fac- Mean 66 years; range 52–81). Inspection of averaged images for tor = 200; single average; FOV = 240 × 200 × 160 mm; acquired these groups suggested regionally specific age-related declines in 18 matrix = 240 × 200. One hundred and sixty slices were reconstructed [ F]FEOBV binding accompanying aging. Selected regions exhibiting to 1-mm isotropic resolution. possible age-related declines were compared with regions that appeared unchanged with aging. Examined regions included the striatal complex, thalamus, primary sensorimotor cortex (BAs 1–35), ante- 2.4 | Analysis of [18F]FEOBV binding rior cingulate cortex (BAs 24,32), hippocampal formation, amygdala, The primary analysis was performed using the average of the co- posterior cingulate cortex (BAs 23,31), cerebellar vermis, occipital cor- registered 180–210 min scan period (termed “late static”). For the tex (BAs 17–19), lateral frontal cortex (BAs 44–47), and medial frontal seven scans that had a slightly different acquisition sequence, we used cortex (BAs 8–10). the average of both the 150–180 and 210–240 time windows, thus matching the median of the scan data used of 195 min. FreeSurfer 2.6 | Statistical analysis (freesurfer-i386-apple-darwin11.4.2-stable6-20170119; https://surfer. 18 nmr.mgh.harvard.edu/fswiki/FreeSurferMethodsCitation; RRID: SCR_ [ F]FEOBV binding was compared within some regions using stan- 001847) was used on each individual's T1-weighted MR to define gray dard parametric methods. Analyses are described in each relevant matter and white matter voxels. The entire co-registered PET sequence Results section. was registered to the MR. The segmented white matter mask from FreeSurfer (1 for WM; 0 elsewhere) was smoothed to PET resolution, 3 | RESULTS and then a threshold of 0.90 applied to the mask, yielding a volume of interest where the partial volume contribution from nonwhite matter regions to the mask was <10% for every voxel. This white matter 3.1 | Overall volume-of-interest was then applied to the PET data and used as a nor- Brain [18F]FEOBV binding was distributed inhomogeneously (Figure 1; malizing (intensity scaling) factor for the late static scan. Accordingly, Table 1). Binding was highest in the striatal complex, followed by por- [18F]FEOBV binding in each volume of interest was expressed as stan- tions of the thalamus, amygdala and hippocampal formation, some dardized uptake value ratios (SUVRs). We showed previously that this neocortical regions, rostral brainstem, and some regions of the cere- approach is equivalent to kinetic modeling approaches for quantifying bellar cortex. The overall distribution and regional density of [18F] regional brain [18F]FEOBV binding (Albin et al., 2017). FEOBV binding is largely consistent with the distribution of choliner- We used NeuroStat (https://neurostat.neuro.utah.edu/) and Statis- gic terminals described in prior human postmortem and nonhuman tical Parametric Mapping (SPM12; http://www.fil.ion.ucl.ac.uk/spm/; mammal studies (Albin et al., 1994; Barmack, Baughman, & Ecken- RRID: SCR_007037) to nonlinearly warp all PET scans into the Talairach stein, 1992; Barmack, Baughman, Eckenstein, & Shojaku, 1992; de and International Consortium for Brain Mapping atlas spaces, respec- Lacalle et al., 1993, 1994; Emre et al., 1993; Fitzpatrick et al., 1989; tively. This was done to allow the creation of mean images of all sub- Fukushima et al., 2001; Gilmor et al., 1996, 1999; Heckers et al., jects, plus mean images for young and elderly cohorts, and to allow 1992; Ichikawa et al., 1997; Jaarsma et al., 1997; Levey et al., 1987; ALBIN ET AL. 2887

FIGURE 1 [18F]FEOBV binding in human brain. (a) Dorsal to ventral survey of averaged transaxial images for all 29 subjects. Images scaled to peak SUVR of 3.0. Regions with notable [18F]FEOBV binding include the striatal complex, thalamus, amygdala, hippocampal formation, neocortical mantle, mesonpontine junction, and portions of the cerebellum (see text for details). (b) Pseudocolor images of dorsal transaxial levels (identical to top row in [a] above) to illustrate differing [18F]FEOBV binding in different cortical regions. Scaled to peak SUVR of 2.5. These images demonstrate relatively higher [18F]FEOBV binding in primary sensorimotor and anterior cingulate cortices

Mahady et al., 2017; Manaye et al., 1999; Mesulam, 2004; Mesulam medial habenular–interpeduncular nucleus projection or the parabigem- et al., 1989; Mesulam & Geula, 1988; Mesulam, Hersh, et al., 1992; inal nucleus to collicular projection, though the parabigeminal nucleus Mesulam, Mash, et al., 1992; Roghani et al., 1998; Schafer et al., may be included in a region of relatively high [18F]FEOBV binding at 1995; Schäfer et al., 1998; Woolf, 1991; Woolf et al., 1984; Woolf & the mesopontine junction (see below). It was not possible to identify Butcher, 1985, 1986, 1989; Zhang et al., 2016). Our results are con- brainstem motor or autonomic cholinergic nuclei. More detailed sistent as well with prior VAChT ligand human SPECT imaging studies description of the identifiable projections–regions follows. (Kuhl et al., 1994, 1996; Lamare et al., 2013). Our results likely identify the distribution of and quantify the regional density of major cholinergic terminal systems in the human 3.2 | Striatal complex brain; cholinergic projections originating in the BF, striatal cholinergic Consistent with the known very high expression of cholinergic interneurons, ascending and descending projections from the PPN/LDT markers in the striatal complex, striatal complex [18F]FEOBV binding complex, and cholinergic MVC projection neurons. It was not possible was the highest of all brain regions, well above levels expressed in any to clearly identify cholinergic nuclei or terminals associated with the other region. Striatal complex [18F]FEOBV binding was mildly 2888 ALBIN ET AL.

TABLE 1 Bilaterally averaged [18F]FEOBV binding within selected volumes of interest

Region Mean SUVR SD Region Mean SUVR SD BA1 1.80 0.19 BA43 1.84 0.18 BA2 1.77 0.17 BA44 1.78 0.14 BA3 1.84 0.20 BA45 1.70 0.15 BA4 1.92 0.23 BA46 1.68 0.15 BA5 1.74 0.19 BA47 1.81 0.15 BA6 1.87 0.19 Amygdala 3.14 0.39 BA7 1.60 0.13 Hippocampal formation 2.76 0.31 BA8 1.66 0.14 Insula 2.57 0.29 BA9 1.71 0.13 Caudate 7.67 1.31 BA10 1.69 0.14 Dorsal caudate 8.00 1.41 BA11 1.67 0.14 Ventral striatum 7.00 1.29 BA12 1.82 0.18 Putamen 8.89 1.43 BA17 1.53 0.15 Anterior putamen 9.01 1.47 BA18 1.48 0.12 Posterior putamen 8.53 1.40 BA19 1.52 0.12 Thalamus 3.45 0.57 BA20 1.65 0.16 Lateral geniculate nucleus 3.47 0.60 BA21 1.73 0.15 Cerebellar hemisphere cortex 1.80 0.16 BA22 1.89 0.17 Cerebellar vermis cortex 3.08 0.38 BA23 1.68 0.14 Flocculus 2.22 0.33 BA24 2.12 0.23 Midbrain 2.35 0.28 BA25 2.89 0.63 Substantia Nigra 2.01 0.27 BA26 1.87 0.27 Raphe nucleus 2.61 0.41 BA27 2.37 0.57 Dorsal midbrain 2.85 0.30 BA28 2.50 0.39 Pons 2.09 0.25 BA29 1.45 0.14 Rostral pons 2.35 0.28 BA30 1.65 0.23 Caudal Pons and Medulla 2.38 0.25 BA31 1.75 0.14 Medulla 2.05 0.28 BA32 2.01 0.20 BA34 2.71 0.40 BA35 2.20 0.35 heterogeneous. [18F]FEOBV binding was highest in the putamen and surrounding cortices (Figures 1 and 2). Similarly, there was a modest lower in the caudate nucleus (Table 1; paired t test, p < .0001). The anterior to posterior gradient with frontal cortices exhibiting modestly striatal complex was further subdivided in dorsal caudate, ventral stri- higher levels of [18F]FEOBV binding than occipital cortices (Figures 1 atum (nucleus accumbens and ventral caudate), anterior putamen, and and 2). Insular cortex was slightly higher (mean SUVR = 2.57; Table 1) posterior putamen (Table 1). [18F]FEOBV binding was modestly differ- As expected, the hippocampal formation (mean SUVR = 2.76) and ent in all these subregions (one-way ANOVA for correlated samples, amygdala (mean SUVR = 3.14) exhibited the highest [18F]FEOBV p < .0001; all subregions different from each other, Tukey's HSD, binding of any cortical region (Table 1, Figure 3). p < .01). The rank order of striatal complex [18F]FEOBV binding was Anterior Putamen > Posterior Putamen > Dorsal Caudate > Ventral 3.4 | Thalamus Striatum. The thalamus exhibited relatively high but likely inhomogeneous [18F] FEOBV binding (Table 1, Figures 1 and 4). Grossly, thalamic [18F] 3.3 | Cortex FEOBV binding distribution did not exhibit the expected ovoid shape [18F]FEOBV binding was distributed widely and somewhat inhomo- (Figures 1 and 4). In particular, the anterolateral border of thalamic geneously throughout the cortex. Most neocortical regions (BA1–23, [18F]FEOBV binding had a concave shape, suggesting less [18F]FEOBV 26, 29–32, 36–47; Figures 1 and 2) exhibited modest [18F]FEOBV binding in the anterolateral thalamus. In addition, the lateral geniculate binding (mean SUVRs = 1.5–2.0 range; Table 1) with some regions nuclei exhibited high [18F]FEOBV binding and could be easily distin- (BA24, 25, 27, 28, 34, and 35) exhibiting higher levels of [18F]FEOBV guished from the rest of the thalamus (Table 1, Figure 4). This sug- binding (mean SUVRs = 2.0–2.5 range; Table 1). Consistent with the gests that portions of the thalamus more rostral and dorsal to the results of the Brodmann areas based analysis, primary sensorimotor lateral geniculate nucleus, for example, portions of the pulvinar com- and anterior cingulate cortices were readily distinguished from plex, had less [18F]FEOBV binding. In the absence of an established ALBIN ET AL. 2889

FIGURE 2 Neocortical [18F]FEOBV binding. Transaxial images from near the vertex to the level of the striatum. Images scaled to emphasize differences in cortical regions. Consistent with Brodmann area based analysis (Table 1; see text), higher [18F]FEOBV binding in primary sensorimotor (SMC) and anterior cingulate (AC) cortices. There is a modest anterior to posterior gradient with higher frontal cortical than occipital cortical [18F]FEOBV binding parcellation of the thalamus, we divided the thalamus into eight subre- the dorsal regions of the caudal brainstem, possibly within the peria- gions by assuming that the thalamus has ovoid shape with a major queductal tegmentum (Figures 1 and 6). rostral-caudal axis, excluding the lateral geniculate nuclei (see section 2). In general, consistent with the impression that anterolateral 3.7 | Aging effects regions exhibited less [18F]FEOBV binding, the lateral subregions Potential aging effects were assessed initially by comparing younger defined by this analysis had less [18F]FEOBV binding than medial sub- (22–38 years) and older (52–81 years) subjects. These comparisons regions and anterior subregions had less [18F]FEOBV binding than revealed regionally specific changes accompanying aging (Table 3, Fig- posterior regions (Table 2). This was confirmed by statistical compari- ures 7 and 8). Statistical significance of age-related changes was sons of cognate octants (e.g., superior anterolateral octant vs. superior assessed with one-way independent t tests using the Holm– anteromedial octant; superior anterolateral octant vs. superior pos- Bonferroni method to correct for multiple comparisons at a threshold terolateral octant, etc.). All medial octants had significantly higher of p < .05. Significant age-related changes were found in the striatum, [18F]FEOBV binding than cognate lateral octants and all posterior primary sensorimotor cortex, and anterior cingulate cortex with trends octants had significantly higher [18F]FEOBV binding than cognate anterior octants (p < 0.05; paired t tests with Bonferroni correction for multiple comparisons). This approach may not sample the pulvinar complex adequately. As seen in Figure 4B, the lateral geniculate nuclei appear relatively isolated, suggesting that the pulvinar complex, rostral and dorsal to the lateral geniculate, exhibits relatively low [18F]FEOBV binding.

3.5 | Cerebellum

All regions of cerebellar cortex exhibited [18F]FEOBV binding with some regions exhibiting relatively high [18F]FEOBV binding levels (Table 1, Figures 1 and 5). Highest cerebellar cortical [18F]FEOBV binding was found in the vermis (mean SUVR = 3.1) with the cerebellar hemisphere cortices at significantly lower levels (mean SUVR = 1.8). The floccular region exhibited intermediate levels of [18F]FEOBV binding (mean SUVR = 2.2).

3.6 | Brainstem

18 – Significant [ F]FEOBV binding (mean SUVRs in the 2 2.5 range) was FIGURE 3 [18F]FEOBV binding in amygdala and hippocampal present in all regions quantified (Table 1). Precise parcellation of the formation. (a,b) [18F]FEOBV binding. (a) Transaxial averaged image of brainstem is very limited due to the small size of brainstem structures all 29 subjects at level of anterior temporal lobe. (b) Parasagittal such as PPN–LDT complex and cranial nerve nuclei. [18F]FEOBV bind- averaged image of all 29 subjects at level of amygdala (AM) and hippocampal formation (HIP). (c,d); early phase averaged images, ing was particularly concentrated in dorsal brainstem from the caudal representing the relative regional blood–brain ligand transport rate— midbrain to the mid pons (Figures 1 and 6). While difficult to quantify, Regional anatomy. (a,b) show significant [18F]FEOBV binding and there also appeared to be modestly increased [18F]FEOBV binding in easily distinguishable amygdala and hippocampal formation 2890 ALBIN ET AL.

FIGURE 4 Thalamic [18F]FEOBV binding. Pseudocolor images of [18F]FEOBV binding (top row) and relative regional blood–brain ligand transport rate—Regional anatomy (bottom row). (a) Transaxial and parasagittal images at levels of dorsal thalamus (THAL). Note the inverted comma shape of high density of [18F]FEOBV binding in thalamus in the transaxial image compared with the ovoid shape of the thalamus in the transaxial flow image. (b) Transaxial and parasagittal images at levels of the lateral geniculate nuclei (LG). Not the relative isolation of high [18F]FEOBV binding in the lateral geniculate nuclei compared with surrounding regions, particularly the more anterior and dorsal pulvinar complex

TABLE 2 Bilaterally averaged thalamic octant SUVRs towards age-related declines in the thalamus, hippocampal formation,

Octant Mean SUVR SD amygdala, and posterior cingulate (Table 3). There was no evidence of Superior anterior lateral 2.82 0.57 age-related decline in the cerebellar vermis and other cortical regions Superior anterior medial 3.15 0.55 (Table 3). To further explore the relationship between aging and Superior posterior lateral 3.40 0.71 regional [18F]FEOBV binding, age was regressed against [18F]FEOBV Superior posterior medial 3.53 0.70 binding in selected regions (Pearson product–moment correlations; Inferior anterior lateral 2.79 0.53 Figure 8). Strong negative correlations were present in striatum Inferior anterior medial 3.61 0.48 Inferior posterior lateral 3.28 0.74 Inferior posterior medial 4.09 0.61

FIGURE 5 Cerebellar [18F]FEOBV binding. Pseudocolor images of [18F]FEOBV binding (top row) and relative regional blood–brain ligand transport rate—Regional anatomy (bottom row). (a) Transaxial images FIGURE 6 Brainstem [18F]FEOBV binding. Pseudocolor parasagittal at the level of flocculi (FL). Note relatively high [18F]FEOBV binding in images of [18F]FEOBV binding (top) and relative regional blood–brain the floccular region, anterior vermis (VER), and posterior vermis (VER) ligand transport rate—Regional anatomy (bottom) in the midline plane. with less [18F]FEOBV binding in the cerebellar hemispheres. Relatively high [18F]FEOBV binding in the dorsal midbrain and (b) Parasagittal images at the level of the flocculus. (c) Coronal images spanning the mesopontine junction into rostral pons (MP). Increased at the level of the vermis. Again, note relatively high [18F]FEOBV [18F]FEOBV binding within the more dorsal caudal brainstem, possibly binding in the vermis compared with the cerebellar hemispheres around the aqueduct ALBIN ET AL. 2891

TABLE 3 Age-related regional changes in [18F]FEOBV binding. Values are SUVRs

Region Young subject mean (SD) Older subject mean (SD) p Striatal complex 9.16 (1.12) 7.39 (0.96) <.0001* Primary sensorimotor cortex 1.92 (0.15) 1.73 (0.96) .002* Anterior cingulate cortex 2.18 (0.16) 1.97 (0.20) .0023* Thalamus 3.68 (0.43) 3.26 (0.61) .024 Hippocampal formation 2.89 (0.32) 2.66 (0.28) .026 Posterior cingulate cortex 1.76 (0.11) 1.68 (0.14) .062 Amygdala 3.26 (0.36) 3.05 (0.40) .071 Cerebellar vermis 3.16 (0.49) 3.02 (0.26) .165 Occipital cortex 1.52 (0.08) 1.50 (0.15) .325 Medial frontal cortex 1.69 (0.08) 1.66 (0.16) .280 Lateral frontal cortex 1.76 (0.08) 1.73 (0.17) .300

*Significant after Holm–Bonferroni correction.

FIGURE 7 Effects of aging on [18F]FEOBV binding. (a) Averaged transaxial pseudocolor images scaled at a peak of 10.0. a1 and a3—Young normal subjects. a2 and a4—Older normal subjects. Note declines in [18F]FEOBV binding in primary sensorimotor and anterior cingulate cortices in older normal subjects. (B) Averaged transaxial pseudocolor images scaled at a peak of 3.0 and at the levels of the striatal complex. b1—Young normal subjects. b2—Older normal subjects. Note decline in striatal complex [18F]FEOBV binding in older subjects 2892 ALBIN ET AL.

striatal complex. This result is consistent with a large body of data from studies in many species indicating uniquely high expression of cholinergic system markers and terminals in the striatal complex. There are at least two, and possibly three, sources of striatal cholinergic terminals. While striatal cholinergic interneurons comprise only a small fraction (~2–3%) of the total striatal neuron population (Bernácer, Prensa, & Giménez-Amaya, 2007; Gonzales & Smith, 2015; Lecumberri, Lopez-Janeiro, Corral-Domenge, & Bernacer, 2018) they give rise to dense axonal-terminal arborizations and play an important role in the integration of cortico- and thalamo-striate inputs with dopaminergic functions. The striatal complex also receives innervation from mesopontine (PPN–LDT) cholinergic neurons (Dautan et al., 2014; Dautan, Hacioglu Bay, Bolam, Gerdjikov, & Mena-Segovia, 2016; Woolf, 1991), presumably collaterals of neurons innervating the thalamus and other targets of the PPN-LDT complex. Mesulam, Hersh, et al. (1992) suggested that the human striatal complex receives some cholinergic afferents from the basal forebrain complex, a projection that is absent in rodents (Dautan et al., 2016). The rela- 18 FIGURE 8 Correlations between age and regional [ F]FEOBV tive proportion of striatal cholinergic interneuron terminals and extrin- binding. Pearson product–moment correlations between subject age sic afferent terminals is unknown. It is generally assumed that striatal and regional [18F]FEOBV binding in striatum, primary sensorimotor cholinergic interneuron terminals are much more abundant than cortex, anterior cingulate cortex, and thalamus. All correlations significant after Bonferroni correction at a threshold of 0.05 extrinsic cholinergic afferent terminals. Janickova et al. (2017) geneti- cally ablated VAChT in murine PPN-LDT neurons. As measured by (r = .68; p < .00003; decline of ~4% per decade), primary sensorimotor western immunoblots, VAChT protein expression diminished markedly cortex (r = .52; p < .003; decline of ~2.5% per decade), and anterior cingu- in the brainstem and thalamus and was unchanged in hippocampal late cortex (r = .55; p < .001; decline of ~2.5% per decade), and thalamus formation, neocortex, and striatum. Genetic or toxic ablation of stria- (r = 0.42; p < .0125; decline of ~2.5% per decade). All correlations signifi- tal cholinergic interneurons results in marked reductions in striatal cant after Bonferroni correction at a threshold of 0.05. An interesting cholinergic terminal markers (Aoki, Liu, Zucca, Zucca, & Wickens, example of regionally specific decrease in [18F]FEOBV binding was found 2015; Laplante, Lappi, & Sullivan, 2011; Martos, Braz, Beccaria, in the primary sensorimotor cortex. In younger subjects, this region was Murer, & Belforte, 2017; Pappas et al., 2015). These results are con- easily distinguished from surrounding cortices by higher [18F]FEOBV sistent with the traditional view that striatal extrinsic cholinergic affer- binding (Figure 7). In older subjects, primary sensorimotor cortex [18F] ents contribute only modestly to the total density of striatal FEOBV binding appeared more similar to surrounding cortices (Figure 7). cholinergic terminals. This is not to say that extrinsic cholinergic striatal afferents are functionally unimportant. Our data suggest modest inhomogeneity of striatal cholinergic 4 | DISCUSSION terminal density. [18F]FEOBV binding was higher in the putamen than the caudate. Further subdividing these regions, we found highest [18F] VAChT is thought to be uniquely expressed by cholinergic neurons. The FEOBV binding within in the anterior putamen and lowest in the ven- VAChT gene is part of a distinctive cholinergic “operon,” encoded within tral striatum. In a stereological analysis of human postmortem mate- the first exon of the Choline Acetyltransferase gene (Eiden, 1998). FEOBV rial, Bernácer et al. (2007) described the inhomogeneous distribution is a high affinity and specific ligand for VAChT (Mulholland et al., 1998). In of striatal cholinergic interneuron perikarya with higher density in the general, our findings correlate well with prior studies of brain cholinergic caudate nucleus than putamen. This is somewhat discordant with our systems in humans and other mammals based on ChAT and high-affinity results. While their definition of subregions is somewhat different choline transporter immunohistochemistry (Kus et al., 2003; and see from ours, their definition of caudate and putamen approximate our below). Our results also correlate well with nonhuman animal and limited definition of dorsal caudate nucleus and putamen. The differences human studies of the distribution of brain VAChT immunoreactivity between our results and those of Bernácer et al. (2007) cautions (Arvidsson et al., 1997; Gilmor et al., 1996, 1999; Ichikawa et al., 1997; against simple equation of perikaryal density and terminal density. Roghani et al., 1998; Schafer et al., 1995; Schäfer et al., 1998). Our results Mesulam, Mash, et al. (1992) suggested that human cholinergic basal also correspond very well with prior limited human PET imaging with [18F] forebrain projections to the striatal complex more heavily innervate FEOBV (Aghourian et al., 2017; Albin et al., 2017; Petrou et al., 2014). the putamen.

4.1 | Striatal complex 4.2 | Cortex

The striatal complex exhibited the highest regional [18F]FEOBV bind- The cortical mantle receives its cholinergic inputs from cholinergic ing, indicating a very high density of cholinergic terminals within the neurons of the basal forebrain complex (Ch1–Ch4 in the ALBIN ET AL. 2893 nomenclature of Mesulam, Mufson, Wainer, & Levey, 1983). These humans, basal forebrain cholinergic afferents probably innervate more cholinergic projections are known to participate in several important anterior and medial thalamic nuclei, including the central medial and cognitive processes including attention, executive functions, and central lateral nuclei, and portions of the mediodorsal nucleus forms of memory (Ballinger et al., 2016; Hasselmo & Sarter, 2011; (Heckers et al., 1992). Prado et al., 2017). Our results are broadly consistent with data from Cholinergic afferents to the thalamus are implicated in several older analyses of cholinergic terminal markers in human postmortem important phenomena, notably the regulation of consciousness and specimens (de Lacalle et al., 1994; Emre et al., 1993; Mesulam, 2004, sleep, though the abundance of cholinergic terminals and receptors 2013; Mesulam & Geula, 1988; Mesulam, Hersh, et al., 1992). Consis- within the thalamus suggests cholinergic neurotransmission may mod- tent with these prior descriptions, highest [18F]FEOBV binding was ulate multiple aspects of thalamic function. Recent work, for example, found in amygdala and hippocampal formation, followed by paralimbic from our group suggests that thalamic cholinergic neurotransmission cortices, including the insular cortex, perirhinal cortex (BA35), entorhi- contributes to “bottom-up” detection of salient sensory cues (Kim, nal cortex (BA34, BA28), parahippocampal gyrus (BA27), and portions Müller, Bohnen, Sarter, & Lustig, 2017). of the cingulate cortex (BA25, BA24). Regional binding density is probably not the only important difference in cholinergic innervation 4.4 | Cerebellum between cortical regions. Mesulam, Hersh, et al. (1992) also described Prior work in several species suggests significant cholinergic innerva- varying laminar distributions of cholinergic terminals between neocor- tion of some cerebellar regions (Barmack, Baughman, & Eckenstein, tical regions. Detecting such differences is beyond the resolution of 1992; Barmack, Baughman, Eckenstein, & Shojaku, 1992; de Lacalle PET imaging. Similarly, prior studies indicate that the basolateral et al., 1993; Fukushima et al., 2001; Jaarsma et al., 1997; Manaye amygdala exhibits the highest density of cholinergic terminals with et al., 1999; Zhang et al., 2016). There is no evidence of cholinergic other amygdala nuclei exhibiting significantly less cholinergic innerva- neurons within the cerebellum in most species studied. Cholinergic tion (Emre et al., 1993). terminals were detected throughout the cerebellar cortex, though most studies describe these as relatively sparse in most cerebellar cor- 4.3 | Thalamus tical regions. [18F]FEOBV PET is likely to convey a more faithful repre- Analysis of the thalamus was limited by lack of an accepted parcella- sentation of cerebellar cortical innervation than AChase PET methods, tion method for thalamic subnuclei. Nonetheless, our data suggest sig- which show uniform and high cerebellar cortical tracer retention. Bar- nificant inhomogeneity for [18F]FEOBV binding within the thalamus. mack and colleagues (Barmack, Baughman, & Eckenstein, 1992; Bar- The lateral geniculate nuclei exhibited significant [18F]FEOBV binding. mack, Baughman, Eckenstein, & Shojaku, 1992) described relatively Visual inspection suggests less [18F]FEOBV binding in anterior, lateral high levels of cholinergic innervation of some cerebellar cortical and some posterior (pulvinar) thalamic subregions. These conclusions regions in rat, rabbit, car, and nonhuman primate brain. These included are consistent with analysis based on the arbitrary division of the thal- the uvula-nodulus (lobules 9 and 10), the floccular region, and the amus into eight equivalent octants. Lateral regions exhibited less [18F] anterior vermis (lobules 1 and 2) (Barmack, Baughman, & Eckenstein, FEOBV binding than medial regions and anterior regions less [18F] 1992; Barmack, Baughman, Eckenstein, & Shojaku, 1992). The pre- FEOBV binding than posterior regions. This crude parcellation dominant source of these terminals appears to be cholinergic neurons approach may not adequately assess [18F]FEOBV binding in the pulvi- of the vestibular complex. Jaarsma et al. (1997) also described cholin- nar complex, which likely also exhibits lower [18F]FEOBV binding. ergic afferents to the cerebellar cortex from numerous brainstem Our results are consistent with the work of Heckers et al. (1992) populations, though MVC cholinergic neurons were felt to be the pre- assessing cholinergic terminal density with ChAT immunohistochemis- dominant source of cerebellar cortical afferents. de Lacalle try in human postmortem specimens. Heckers et al. (1992) documen- et al. (1993) described a ChAT-immunoreactive subpopulation of cere- ted ChAT immunoreactive fibers throughout the human thalamus bellar cortical Golgi interneurons, particularly in the vermis, uvula- with the greatest density of ChAT-immunoreactive fibers in the lateral nodulus, and flocculus, in human postmortem specimens. geniculate nuclei, some intralaminar nuclei, the reunions nucleus, the We find uniformly high [18F]FEOBV binding throughout the ver- anterodorsal nucleus, the reticular nucleus, and some portions of the mis, different from the patterns described in other mammals. The orig- mediodorsal nucleus. Lowest density of ChAT-immunoreactive fibers inating neurons of this extensive cerebellar vermis cholinergic was seen in the pulvinar and medial geniculate nuclei. As Heckers innervation are unknown, though it is possible that it is an extension et al. (1992) point out the human subregional distribution of thalamic of the vestibulocerebellar cholinergic innervation documented in cholinergic terminals is likely distinct from that found in the limited other mammals. Alternatively, the pattern of cerebellar [18F]FEOBV number of other species studied (Fitzpatrick et al., 1989; Levey binding may reflect the distribution of cholinergic Golgi interneurons et al., 1987). described by de Lacalle et al. (1993). The vestibular system plays a The major cholinergic innervation of thalamus originates in the critical role in the maintenance of normal upright posture and gait, as PPN–LDT complex with some contribution from the basal forebrain. does the cerebellar vermis. A plausible speculation is that extended PPN–LDT cholinergic efferent projections generally ramify widely and vestibular cholinergic innervation of the cerebellar vermis is a corre- innervate several target regions (Mena-Segovia & Bolam, 2017), late of bipedal locomotion in humans. Seidel et al. (2015) document though Holmstrand and Sesack (2011) suggest some specialization of α-synuclein aggregate pathology in the MVC nuclei of postmortem cholinergic PPN–LDT afferents targeting the anterior thalamus. In specimens from individuals with Parkinson disease and Lewy body 2894 ALBIN ET AL. dementia. It is possible that vestibular dysfunction contributes to the neurons in humans. Zhang, Sampogna, Morales, & Chase (2005) stud- dopamine replacement therapy refractory gait and balance deficits of ied PPN–LDT neurons in aged cats. There was no evidence of PPN– advanced Parkinson disease. LDT cholinergic perikarya loss but there was evidence of neuronal Another previously documented site of cholinergic innervation of atrophy and dendritic simplification. Prior work suggests age-related the cerebellum is the deep cerebellar nuclei. Prior tract-tracing and changes in BF neurons. MRI morphometry studies indicate age-related functional studies in rodents indicate projections from the PPN-LDT loss of BF nuclear volumes (Grothe, Heinsen, & Teipel, 2012; Hanyu complex to the deep cerebellar nuclei with MRI tractography studies et al., 2002). These results are consistent with our data suggesting supporting the existence of this projection in humans (Aravamuthan, age-related loss of BF corticopetal terminals. Our data suggest that Muthusamy, Stein, Aziz, & Johansen-Berg, 2007; Vitale et al., 2016). this is not a uniform process as we found cortical regions (primary These terminals are presumably too sparse to be detected by [18F] sensorimotor cortex, anterior cingulate cortices) with significant FEOBV PET. declines in [18F]FEOBV binding and others with little or no age-related change. As subpopulations of BF neurons are thought to project to 4.5 | Brainstem relatively restricted cortical regions (Gielow & Zaborszky, 2017), our results suggest inhomogeneous age-related changes within the BF Cholinergic neurons within the brainstem include oculomotor (CrN III, complex. Prior SPECT studies of human VAChT distribution also sug- IV, and VI nuclei), motor (motor nucleus of CrN V, and CrN VII, CrN X, gested age-related decline in BF corticopetal terminals but lacked the XI, and CrN XII nuclei), and autonomic nuclei (Vagal Nucleus), and precision to resolve changes in individual neocortical regions (Kuhl some MVC neurons. Most of these nuclei are beyond the resolution et al., 1996). of PET imaging. The largest collection of brainstem cholinergic neurons lies within the PPN-LDT complex (Mesulam et al., 1983, Chap- 4.7 | General conclusion ter 5 and 6). In immunohistochemical studies of postmortem human specimens, Mesulam et al. (1989) described the PPN–LDT complex as [18F]FEOBV PET delineates regional cholinergic terminals well in stretching from the dorsal mid-mesencephalon to dorsal mid-pons. human brains. The general pattern of [18F]FEOBV binding is highly The distribution of cholinergic perikarya did not observe specific consistent with the known organization of major cholinergic systems nuclear borders nor was it delimited by fiber tracts. Cholinergic peri- in mammalian brain. Our results indicate distinctive organization of karya were distributed at varying densities throughout this volume of cholinergic terminals in the human neocortex, cerebellar cortex, and the brainstem. Similar results were obtained by Manaye et al. (1999). the human thalamus. Our results indicate the presence of age-related In rodent PPN–LDT, there is a rostrocaudal gradient of cholinergic and region-specific changes in cholinergic terminal density. [18F] perikarya density with the highest density of these neurons in the FEOBV PET is likely to be a useful method for studying disease states, caudal PPN–LDT complex (Mena-Segovia & Bolam, 2017). The distri- including several neurodegenerative disorders and psychiatric bution of cholinergic perikarya described in humans by Mesulam disorders. et al. (1989) coincides well with the distribution of relatively high [18F] FEOBV binding we find spanning the dorsal mesopontine junction. This area of relatively high brainstem [18F]FEOBV binding may also ACKNOWLEDGMENTS include the cholinergic parabigeminal nucleus (Mesulam et al., 1983, The authors thank our research participants and the technical staff of Chapter 8), adjacent and lateral to the PPN-LDT complex in the the University of Michigan PET Center. This study was supported by upper pons. P01 NS015655, P50 NS091856, R21 NS088302, and the Michael We found significant [18F]FEOBV binding throughout the brain- J. Fox Foundation. stem, most likely the binding to the terminals of PPN–LDT neurons that innervate multiple targets throughout the brainstem (Woolf & ORCID Butcher, 1989). While our analysis is limited by the resolution of PET, Roger L. Albin https://orcid.org/0000-0002-0629-608X prior studies in other species do not describe marked inhomogeneity of brainstem cholinergic terminals. It was possible to tentatively iden- REFERENCES tify what might be a column of relatively increased [18F]FEOBV bind- Aghourian, M., Legault-Denis, C., Soucy, J. P., Rosa-Neto, P., Gauthier, S., ing in the dorsal caudal brainstem surrounding the aqueduct Kostikov, A., … Bédard, M. A. (2017). Quantification of brain choliner- (Figures 1 and Figure 5). gic denervation in Alzheimer's disease using PET imaging with [18F]-FEOBV. Molecular Psychiatry, 22, 1531–1538. Albin, R. L., Koeppe, R. A., Bohnen, N. I., Wernette, K., Kilbourn, M. A., & 4.6 | Aging effects Frey, K. A. (2008). Spared caudal brainstem SERT binding in early Par- kinson's disease. Journal of Cerebral Blood Flow and Metabolism, 28, Our results indicate significant and regionally specific declines in [18F] 441–444. FEOBV binding. Striatal complex changes are consistent with age- Albin, R. L., Minderovic, C., & Koeppe, R. A. (2017). Normal striatal vesicu- related declines in striatal cholinergic interneuron terminal density. lar acetylcholine transporter expression in Tourette syndrome. eNeuro, There was a suggestion of reduced thalamic [18F]FEOBV binding, con- 4, 017817. Albin, R. L., Morgan, M. M., Higgins, D. S., & Frey, K. A. (1994). Autoradio- – sistent with age-related decrease in PPN LDT complex cholinergic graphic quantification of muscarinic cholinergic synaptic markers in terminals. Ransmayr et al. (2000) described age-related loss of PPN bat, shrew, and rat brain. Neurochemical Research, 19, 581–589. ALBIN ET AL. 2895

Anthenelli, R. M., Benowitz, N. L., West, R., St. Aubin, L., McRae, T., Fitzpatrick, D., Diamond, I. T., & Raczkowski, D. (1989). Cholinergic and Lawrence, D., … Evins, A. E. (2016). Neuropsychiatric safety and effi- monoaminergic innervation of the cat's thalamus: Comparison of the cacy of varenicline, bupropion, and nicotine patch in smokers with and lateral geniculate nucleus with other principal sensory nuclei. Journal of without psychiatric disorders (EAGLES): A double-blind, randomised, Comparative Neurology, 288, 647–675. placebo-controlled clinical trial. Lancet, 387, 2507–2520. Fukushima, M., Kitahara, T., Takeda, N., Saika, T., Uno, A., & Kubo, T. Aoki, S., Liu, A. W., Zucca, A., Zucca, S., & Wickens, J. R. (2015). Role of (2001). Role of cholinergic mossy fibers in medial vestibular and prepo- striatal cholinergic interneurons in set-shifting in the rat. Journal of situs hypoglossal nuclei in vestibular compensation. Neuroscience, 102, Neuroscience, 35, 9424–9431. 159–166. Aravamuthan, B. R., Muthusamy, K. A., Stein, J. F., Aziz, T. Z., & Gielow, M. R., & Zaborszky, L. (2017). The input-output relationship of the Johansen-Berg, H. (2007). Topography of cortical and subcortical con- cholinergic basal forebrain. Cell Reports, 18, 1817–1830. nections of the human pedunculopontine and subthalamic nuclei. Neu- Gilman, S., Koeppe, R. A., Nan, B., Wang, C. N., Wang, X., Junck, L., … roImage, 37, 694–705. Bhaumik, A. (2010). Cerebral cortical and subcortical cholinergic defi- Arvidsson, U., Riedl, M., Elde, R., & Meister, B. (1997). Vesicular acetylcho- cits in parkinsonian syndromes. Neurology, 74, 1416–1423. line transporter (VAChT) protein: A novel and unique marker for cho- Gilmor, M. L., Erickson, J. D., Varoqui, H., Hersh, L. B., Bennett, D. A., linergic neurons in the central and peripheral nervous systems. Journal Cochran, E. J., … Levey, A. I. (1999). Preservation of nucleus basalis of Comparative Neurology, 378, 454–467. neurons containing choline acetyltransferase and the vesicular acetyl- Ballinger, E. C., Ananth, M., Talmage, D. A., & Role, L. W. (2016). Basal choline transporter in the elderly with mild cognitive impairment and forebrain cholinergic circuits and signaling in cognition and cognitive early Alzheimer's disease. Journal of Comparative Neurology, 411, decline. Neuron, 91, 1199–1218. 693–704. Barmack, N. H., Baughman, R. W., & Eckenstein, F. P. (1992). Cholinergic Gilmor, M. L., Nash, N. R., Roghani, A., Edwards, R. H., Yi, H., innervation of the cerebellum of rat, rabbit, cat, and monkey as Hersch, S. M., & Levey, A. . I. (1996). Expression of the putative vesicu- revealed by choline acetyltransferase activity and immunohistochemis- lar acetylcholine transporter in rat brain and localization in cholinergic try. Journal of Comparative Neurology, 317, 233–249. synaptic vesicles. Journal of Neuroscience, 16, 2179–2190. Barmack, N. H., Baughman, R. W., Eckenstein, F. P., & Shojaku, H. (1992). Gonzales, K. K., & Smith, Y. (2015). Cholinergic interneurons in the dorsal Secondary vestibular cholinergic projection to the cerebellum of rabbit and ventral striatum: Anatomical and functional considerations in nor- and rat as revealed by choline acetyltransferase immunohistochemis- mal and diseased conditions. Annals of the New York Academy of Sci- try, retrograde and orthograde tracers. Journal of Comparative Neurol- ences, 1349,1–45. ogy, 317, 250–270. Greenfield, S. A. (1991). A noncholinergic action of acetylcholinesterase Bernácer, J., Prensa, L., & Giménez-Amaya, J. M. (2007). Cholinergic inter- (AChE) in the brain: From neuronal secretion to the generation of neurons are differentially distributed in the human striatum. PLoS One, movement. Cellular and Molecular Neurobiology, 11,55–77. 2, e1174. Grothe, M., Heinsen, H., & Teipel, S. J. (2012). Atrophy of the cholinergic Bohnen, N. I., Albin, R. L., Koeppe, R. A., Wernette, K. A., Kilbourn, M. R., basal forebrain over the adult age range and in early stages of Alzhei- Minoshima, S., & Frey, K. A. (2006). Positron emission tomography of mer's disease. Biological Psychiatry, 71, 805–813. monoaminergic vesicular binding in aging and Parkinson disease. Jour- Gut, N. K., & Winn, P. (2016). The pedunculopontine tegmental nucleus-a nal of Cerebral Blood Flow and Metabolism, 26, 1198–1212. functional hypothesis from the comparative literature. Movement Dis- Bohnen, N. I., Müller, M. L., Kotagal, V., Koeppe, R. A., Kilbourn, M. R., orders, 31, 615–624. Gilman, S., … Frey, K. A. (2012). Heterogeneity of cholinergic denerva- Hanyu, H., Asano, T., Sakurai, H., Tanaka, Y., Takasaki, M., & Abe, K. tion in Parkinson's disease without dementia. Journal of Cerebral Blood (2002). MR analysis of the substantia innominata in normal aging, Alz- Flow and Metabolism, 32, 1609–1617. heimer disease, and other types of dementia. American Journal of Neu- Chung, K. A., Lobb, B. M., Nutt, J. G., & Horak, F. B. (2010). Effects of a roradiology, 23,27–32. central cholinesterase inhibitor on reducing falls in Parkinson disease. Hasselmo, M. E., & Sarter, M. (2011). Modes and models of forebrain cho- Neurology, 75, 1263–1269. linergic neuromodulation of cognition. Neuropsychopharmacology, 36, Dautan, D., Hacioglu Bay, H., Bolam, J. P., Gerdjikov, T. V., & 52–73. Mena-Segovia, J. (2016). Extrinsic sources of cholinergic innervation of Heckers, S., Geula, C., & Mesulam, M. M. (1992). Cholinergic innervation the striatal complex: A whole-brain mapping analysis. Frontiers in Neu- of the human thalamus: Dual origin and differential nuclear distribu- roanatomy, 10,1. tion. Journal of Comparative Neurology, 325,68–82. Dautan, D., Huerta-Ocampo, I., Witten, I. B., Deisseroth, K., Bolam, J. P., Henderson, E. J., Lord, S. R., Brodie, M. A., Gaunt, D. M., Lawrence, A. D., Gerdjikov, T., & Mena-Segovia, J. (2014). A major external source of Close, J. C., … Ben-Shlomo, Y. (2016). Rivastigmine for gait stability in cholinergic innervation of the striatum and nucleus accumbens origi- patients with Parkinson's disease (ReSPonD): A randomised, nates in the brainstem. Journal of Neuroscience, 34, 4509–4518. double-blind, placebo-controlled, phase 2 trial. Lancet Neurology, 15, de Lacalle, S., Hersh, L. B., & Saper, C. B. (1993). Cholinergic innervation of 249–258. the human cerebellum. Journal of Comparative Neurology, 328, Hirsch, E. C., Graybiel, A. M., Duyckaerts, C., & Javoy-Agid, F. (1987). Neu- 364–376. ronal loss in the pedunculopontine tegmental nucleus in Parkinson dis- de Lacalle, S., Lim, C., Sobreviela, T., Mufson, E. J., Hersh, L. B., & ease and in progressive supranuclear palsy. Proceedings of the National Saper, C. B. (1994). Cholinergic innervation in the human hippocampal Academy of Sciences of the United States of America, 84, 5976–5980. formation including the entorhinal cortex. Journal of Comparative Neu- Holmstrand, E. C., & Sesack, S. R. (2011). Projections from the rat pedun- rology, 345, 321–344. culopontine and laterodorsal tegmental nuclei to the anterior thalamus Deffains, M., & Bergman, H. (2015). Striatal cholinergic interneurons and and ventral tegmental area arise from largely separate populations of cortico-striatal synaptic plasticity in health and disease. Movement Dis- neurons. Brain Structure and Function, 216, 331–345. orders, 30, 1014–1025. Ichikawa, T., Ajiki, K., Matsuura, J., & Misawa, H. (1997). Localization of DeKosky, S. T., Harbaugh, R. E., Schmitt, F. A., Bakay, R. A., Chui, H. C., two cholinergic markers, choline acetyltransferase and vesicular acetyl- Knopman, D. S., … Markesbery, W. R. (1992). Cortical biopsy in Alzhei- choline transporter in the central nervous system of the rat: In situ mer's disease: Diagnostic accuracy and neurochemical, neuropathologi- hybridization histochemistry and immunohistochemistry. Journal of cal, and cognitive correlations. Intraventricular Bethanecol Study Chemical Neuroanatomy, 13,23–39. Group. Annals of Neurology, 32, 625–632. Iyo, M., Namba, H., Fukushi, K., Shinotoh, H., Nagatsuka, S., Suhara, T., … Eiden, L. E. (1998). The cholinergic gene locus. Journal of Neurochemistry, Irie, T. (1997). Measurement of acetylcholinesterase by positron emis- 70, 2227–2240. sion tomography in the brains of healthy controls and patients with Emre, M., Heckers, S., Mash, D. C., Geula, C., & Mesulam, M. M. (1993). Alzheimer's disease. Lancet, 349, 1805–1809. Cholinergic innervation of the amygdaloid complex in the human brain Jaarsma, D., Ruigrok, T. J., Caffé, R., Cozzari, C., Levey, A. I., and its alterations in old age and Alzheimer's disease. Journal of Com- Mugnaini, E., & Voogd, J. (1997). Cholinergic innervation and receptors parative Neurology, 336, 117–134. in the cerebellum. Progress in Brain Research, 114,67–96. 2896 ALBIN ET AL.

Janickova, H., Rosborough, K., Al-Onaizi, M., Kljakic, O., Guzman, M. S., laterodorsal tegmental nuclei and some cytochemical comparisons to Gros, R., … Prado, V. F. (2017). Deletion of the vesicular acetylcholine forebrain cholinergic neurons. Journal of Comparative Neurology, 283, transporter from pedunculopontine/laterodorsal tegmental neurons 611–633. modifies gait. Journal of Neurochemistry, 140, 787–798. Mesulam, M. M., Hersh, L. B., Mash, D. C., & Geula, C. (1992). Differential Kim, K., Müller, M. L. T. M., Bohnen, N. I., Sarter, M., & Lustig, C. (2017). cholinergic innervation within functional subdivisions of the human Thalamic cholinergic innervation makes a specific bottom-up contribu- cerebral cortex: A choline acetyltransferase study. Journal of Compara- tion to signal detection: Evidence from Parkinson's disease patients tive Neurology, 318, 316–328. with defined cholinergic losses. NeuroImage, 149, 295–304. Mesulam, M. M., Mash, D., Hersh, L., Bothwell, M., & Geula, C. (1992). Kucinski, A., de Jong, I. E., & Sarter, M. (2017). Reducing falls in Parkinson's Cholinergic innervation of the human striatum, globus pallidus, subtha- disease: Interactions between donepezil and the 5-HT6 receptor antag- lamic nucleus, substantia nigra, and red nucleus. Journal of Comparative onist idalopirdine on falls in a rat model of impaired cognitive control of Neurology, 323, 252–268. complex movements. European Journal of Neuroscience, 45,217–231. Mesulam, M. M., Mufson, E. J., Wainer, B. H., & Levey, A. I. (1983). Central Kuhl, D. E., Koeppe, R. A., Fessler, J. A., Minoshima, S., Ackermann, R. J., cholinergic pathways in the rat: An overview based on an alternative Carey, J. E., … Wieland, D. M. (1994). In vivo mapping of cholinergic nomenclature (Ch1-Ch6). Neuroscience, 10, 1185–1201. neurons in the human brain using SPECT and IBVM. Journal of Nuclear Mufson, E. J., Ginsberg, S. D., Ikonomovic, M. D., & DeKosky, S. T. (2003). Medicine, 35, 405–410. Human cholinergic basal forebrain: Chemoanatomy and neurologic Kuhl, D. E., Koeppe, R. A., Minoshima, S., Snyder, S. E., Ficaro, E. P., dysfunction. Journal of Chemical Neuroanatomy, 26, 233–242. Foster, N. L., … Kilbourn, M. R. (1999). In vivo mapping of cerebral ace- Mulholland, G. K., Wieland, D. M., Kilbourn, M. R., Frey, K. A., tylcholinesterase activity in aging and Alzheimer's disease. Neurology, Sherman, P. S., Carey, J. E., & Kuhl, D. E. (1998). [18F]fluoroethoxy-- 52, 691–699. benzovesamicol, a PET radiotracer for the vesicular acetylcholine Kuhl, D. E., Minoshima, S., Fessler, J. A., Frey, K. A., Foster, N. L., transporter and cholinergic synapses. Synapse, 30, 263–274. Ficaro, E. P., … Koeppe, R. A. (1996). In vivo mapping of cholinergic ter- Pappas, S. S., Darr, K., Holley, S. M., Cepeda, C., Mabrouk, O. S., minals in normal aging, Alzheimer's disease, and Parkinson's disease. Wong, J. M., … Dauer, W. T. (2015). Forebrain deletion of the dystonia Annals of Neurolory, 40, 399–410. protein torsinA causes dystonic-like movements and loss of striatal Kus, L., Borys, E., Ping Chu, Y., Ferguson, S. M., Blakely, R. D., cholinergic neurons. eLife, 4, e08352. Emborg, M. E., … Mufson, E. J. (2003). Distribution of high affinity cho- Petrou, M., Frey, K. A., Kilbourn, M. R., Scott, P. J., Raffel, D. M., line transporter immunoreactivity in the primate central nervous sys- Bohnen, N. I., … Koeppe, R. A. (2014). In vivo imaging of human cholin- tem. Journal of Comparative Neurology, 463, 341–357. ergic nerve terminals with (−)-5-(18)F-fluoroethoxybenzovesamicol: Lamare, F., Mazere, J., Attila, M., Mayo, W., De Clermont-Gallerande, H., Biodistribution, dosimetry, and tracer kinetic analyses. Journal of Meissner, W., … Allard, M. (2013). Improvement of in vivo quantifica- Nuclear Medicine, 55, 396–404. tion of [123I]-Iodobenzovesamicol in single-photon emission com- Pienaar, I. S., Vernon, A., & Winn, P. (2016). The cellular diversity of the puted tomography/computed tomography using anatomic image to Pedunculopontine nucleus: Relevance to behavior in health and brain atlas nonrigid registration. Molecular Imaging, 12, 288–299. aspects of Parkinson's disease. The Neuroscientist, 23, 415–431. Laplante, F., Lappi, D. A., & Sullivan, R. M. (2011). Cholinergic depletion in Prado, V. F., Janickova, H., Al-Onaizi, M. A., & Prado, M. A. (2017). Cholin- the nucleus accumbens: Effects on amphetamine response and senso- ergic circuits in cognitive flexibility. Neuroscience, 345, 130–141. rimotor gating. Progress in Neuro-Psychopharmacology & Biological Psy- Prado, V. F., Roy, A., Kolisnyk, B., Gros, R., & Prado, M. A. (2013). Regula- chiatry, 35, 501–509. tion of cholinergic activity by the vesicular acetylcholine transporter. Lecumberri, A., Lopez-Janeiro, A., Corral-Domenge, C., & Bernacer, J. Biochemical Journal, 450, 265–274. (2018). Neuronal density and proportion of interneurons in the asso- Ransmayr, G., Faucheux, B., Nowakowski, C., Kubis, N., Federspiel, S., ciative, sensorimotor and limbic human striatum. Brain Structure & Kaufmann, W., … Hirsch, E. C. (2000). Age-related changes of neuronal Function, 223, 1615–1625. counts in the human pedunculopontine nucleus. Neuroscience Letters, Levey, A. I., Hallanger, A. E., & Wainer, B. H. (1987). Choline acetyltransfer- 288, 195–198. ase immunoreactivity in the rat thalamus. Journal of Comparative Neu- Roghani, A., Shirzadi, A., Butcher, L. L., & Edwards, R. H. (1998). Distribu- rology, 257, 317–332. tion of the vesicular transporter for acetylcholine in the rat central ner- Li, Z., Yu, Z., Zhang, J., Wang, J., Sun, C., Wang, P., & Zhang, J. (2015). vous system. Neuroscience, 82, 1195–1212. Impact of Rivastigmine on cognitive dysfunction and falling in Parkin- Schäfer, M. K., Eiden, L. E., & Weihe, E. (1998). Cholinergic neurons and termi- son's disease patients. European Neurology, 74,86–91. nal fields revealed by immunohistochemistry for the vesicular acetylcho- Mahady, L. J., Perez, S. E., Emerich, D. F., Wahlberg, L. U., & Mufson, E. J. line transporter. I. Central nervous system. Neuroscience, 84,331–359. (2017). Cholinergic profiles in the Goettingen miniature pig (Sus scrofa Schafer, M. K., Weihe, E., Erickson, J. D., & Eiden, L. E. (1995). Human and domesticus) brain. Journal of Comparative Neurology, 525, 553–573. monkey cholinergic neurons visualized in paraffin-embedded tissues Manaye, K. F., Zweig, R., Wu, D., Hersh, L. B., De Lacalle, S., Saper, C. B., & by immunoreactivity for VAChT, the vesicular acetylcholine trans- German, D. C. (1999). Quantification of cholinergic and select porter. Journal of Molecular Neuroscience, 6, 225–235. non-cholinergic mesopontine neuronal populations in the human brain. Schliebs, R., & Arendt, T. (2011). The cholinergic system in aging and neu- Neuroscience, 89, 759–770. ronal degeneration. Behavioural Brain Research, 221, 555–563. Martos, Y. V., Braz, B. Y., Beccaria, J. P., Murer, M. G., & Belforte, J. E. Seidel, K., Mahlke, J., Siswanto, S., Krüger, R., Heinsen, H., Auburger, G., … (2017). Compulsive social behavior emerges after selective ablation of Rüb, U. (2015). The brainstem pathologies of Parkinson's disease and striatal cholinergic interneurons. Journal of Neuroscience, 37, dementia with Lewy bodies. Brain Pathology, 25, 121–135. 2849–2858. Shao, X., Hoareau, R., Hockley, B. G., Tluczek, L. J., Henderson, B. D., Mena-Segovia, J., & Bolam, J. P. (2017). Rethinking the Pedunculopontine Padgett, H. C., & Scott, P. J. (2011). Highlighting the versatility of the nucleus: From cellular organization to function. Neuron, 94,7–18. Tracerlab synthesis modules. Part 1: Fully automated production of [F] Mesulam, M. M. (2004). The cholinergic innervation of the human cerebral labelled radiopharmaceuticals using a Tracerlab FX(FN). Journal of cortex. Progress in Brain Research, 145,67–78. Labelled Compounds & Radiopharmaceuticals, 54, 292–307. Mesulam, M. M. (2013). Cholinergic circuitry of the human nucleus basalis Snyder, S. E., Gunupudi, N., Sherman, P. S., Butch, E. R., Skaddan, M. B., and its fate in Alzheimer's disease. Journal of Comparative Neurology, Kilbourn, M. R., … Kuhl, D. E. (2001). Radiolabeled cholinesterase sub- 521, 4124–4144. strates: in vitro methods for determining structure-activity relation- Mesulam, M. M., & Geula, C. (1988). Nucleus basalis (Ch4) and cortical ships and identification of a positron emission tomography cholinergic innervation in the human brain: Observations based on the radiopharmaceutical for in vivo measurement of butyrylcholinesterase distribution of acetylcholinesterase and choline acetyltransferase. Jour- activity. Journal of Cerebral Blood Flow & Metabolism, 21, 132–143. nal of Comparative Neurology, 275, 216–240. Tanimura, A., Pancani, T., Lim, S. A. O., Tubert, C., Melendez, A. E., Mesulam, M. M., Geula, C., Bothwell, M. A., & Hersh, L. B. (1989). Human Shen, W., & Surmeier, D. J. (2017). Striatal cholinergic interneurons and reticular formation: Cholinergic neurons of the pedunculopontine and Parkinson's disease. European Journal of Neuroscience, 47,1148–1158. ALBIN ET AL. 2897

Vitale, F., Mattei, C., Capozzo, A., Pietrantoni, I., Mazzone, P., & Scarnati, E. Zhang, J. H., Sampogna, S., Morales, F. R., & Chase, M. H. (2005). Age-re- (2016). Cholinergic excitation from the pedunculopontine tegmental lated changes in cholinergic neurons in the laterodorsal and the pedun- nucleus to the dentate nucleus in the rat. Neuroscience, 317,12–22. culo-pontine tegmental nuclei of cats: a combined light and electron Woolf, N. J. (1991). Cholinergic systems in mammalian brain and spinal microscopic study. Brain Research, 1052,47–55. cord. Progress in Neurobiology, 37, 475–524. Zhang, C., Zhou, P., & Yuan, T. (2016). The cholinergic system in the cere- Woolf, N. J., & Butcher, L. L. (1985). Cholinergic systems in the rat brain: bellum: From structure to function. Reviews in the Neurosciences, 27, II. Projections to the interpeduncular nucleus. Brain Research Bulletin, 769–776. 14,63–83. Woolf, N. J., & Butcher, L. L. (1986). Cholinergic systems in the rat brain: III. Projections from the pontomesencephalic tegmentum to the thalamus, tectum, basal ganglia, and basal forebrain. Brain Research Bulletin, How to cite this article: Albin RL, Bohnen NI, Muller MLTM, 16, 603–637. Woolf, N. J., & Butcher, L. L. (1989). Cholinergic systems in the rat brain: et al. Regional vesicular acetylcholine transporter distribution IV. Descending projections of the pontomesencephalic tegmentum. in human brain: A [18F]fluoroethoxybenzovesamicol positron – Brain Research Bulletin, 23, 519 540. emission tomography study. J Comp Neurol. 2018;526: Woolf, N. J., Eckenstein, F., & Butcher, L. L. (1984). Cholinergic systems in 2884–2897. https://doi.org/10.1002/cne.24541 the rat brain: I. projections to the limbic telencephalon. Brain Research Bulletin, 13, 751–784.